%load_ext pretty_jupyter

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import seaborn.objects as so
import numpy as np
from pretty_jupyter.helpers import matplotlib_fig_to_markdown

sns.set_theme()

%%jmd

# Algorithm
[//]: # (-.- .tabset .tabset-pills .tabset-fade)

Sometimes we want to create tabsets dynamically, meaning we want to generate them based on a code.
Pretty Jupyter supports this by using Jinja Markdown cells.

Algorithm¶

Sometimes we want to create tabsets dynamically, meaning we want to generate them based on a code. Pretty Jupyter supports this by using Jinja Markdown cells.

# create dataframe
data = pd.DataFrame({"a": [1, 2, 3, 4], "b": [4, 3, 2, 1], "c": [1, 4, 2, 3]})

%%jmd

{% for col in data.columns %}

## Tabset for column {{ col }}

{% set obj = plt.subplots() %}
{% set _ = so.Plot(x=np.arange(data.shape[0]), y=data[col]).add(so.Line(marker="o")).label(title="Plot for column '{}'".format(col), y=col).on(obj[1]).plot() %}
{{ matplotlib_fig_to_markdown(obj[0]) }}

{% endfor %}

Tabset for column a¶

Tabset for column b¶

Tabset for column c¶

Each Jinja Markdown cell is internally ran through Jinja templating framework. To read more about Jinja, check out its documentation.

Alternative Method¶

There is an alternative method how to do it. If we don't want to work in jinja, we can use pure python for the cycle. We can use Markdown class for that.

Note that inside Markdown class we cannot use tokens (such as .tabset etc).

.alert|O_O|.alert-warning

This method is, however, sometimes useful, if our loop is more code-heavy than markdown-heavy.

from IPython.display import display, Markdown

for col in data.columns:
    display(Markdown(f"## Tabset for column {col}"))

    fig, ax = plt.subplots()
    so.Plot(x=np.arange(data.shape[0]), y=data[col]).add(so.Line(marker="o")).label(title="Plot for column '{}'".format(col), y=col).on(ax).plot()
    display(Markdown(matplotlib_fig_to_markdown(ax.figure)))